Although the literature on homeless individuals coping with drug abuse and other mental illnesses has given some attention to lack of resources as a risk factor for homelessness, the argument made essentially 1 directional--that costs and availability of affordable housing are beyond the means of the individual, leading directly to becoming homeless. This argument generally treats housing affordability and availability as static constructs. However, housing also goes through changes over time---value and availability vary as the unit ages and the neighborhood changes. The purpose of this grant is to test 2 competing constructs of the role that these "future-homeless" tenants play in housing over time, specifically examining the role played by drug abuse/ dependence and other risk factors in housing stock. The first construct is that future-homeless persons represent tenants of last resort. As housing becomes less desirable, any tenant becomes preferable to none, thus individuals at high risk become acceptable. A corollary of this is that pile-ups of risk factors, such as current addiction, increases the likelihood they inhabit this role. The second explanation is that homeless individuals represent the tenants of choice, occupying housing developed specifically for low-income occupants. Methodolgically, testing these theories requires longitudinal data on housing value and occupancy, as well as complete data on last stable housing and individual-level characteristics for a representative homeless sample. Using a unique data set, Service Use, Needs, Costs, and Consequences of Drug Abuse on homelessness (SUNCODA: R01-DA 10713), we propose to combine the last stable addresses for 400 randomly-sampled homeless with publicly available (but unfortunately low technology) housing stock information, controlling for census-level data for the relevant zip code, in order to test the study hypotheses. The 2 proposed theories have strong implications for current service use models and development of primary prevention efforts. Because current services focus on housing as a primary outcome, finding that multiple risk factors lead to housing that is likely to become unstable represents critical new information for improving current services. Thus, this proposal is responsive to the PA-02-150, Services and Intervention Research with Homeless Persons having Alcohol, Drug Abuse, or Mental Disorders. [unreadable] [unreadable] [unreadable] [unreadable]